Is Android fragmentation taking care of itself?
A few weeks ago, I asked if fragmentation of the Android operating system was preventing world domination. It seemed as though Android's greatest strengths (its openness, customizability, and the laissez-faire approach to applications) were also its greatest weaknesses in terms of making it the platform to beat. Where Apple's One-platform-to-rule-them-all ensured a top-notch user experience but left developers bristling, Android's approach left developers wondering where to put their development efforts.
Now, while device fragmentation is still giving developers pause, it appears that most users at least have access to Android 2.1 and above, removing the limitations imposed by earlier versions that remained common in the wild. According to a report on Google's Android developer site, over 70% of Android devices are running at least Android 2.1 and 2.2 adoption is growing quickly. That still leaves almost 30% of users stuck in Android 1.5 and 1.6, most of whom have no option to upgrade because of carrier or device manufacturer restrictions, but the following curves shown on the developer site suggest that usage is dropping off fast enough that developers can essentially ignore the older operating systems:
What they can't ignore is the increasingly wide variety of devices that run Android. Tablets, televisions, small low-end smartphones, large high-end phones, and even the occasional netbook are turning up with Android under the hood. While desktop developers are well accustomed to writing software for a windowed environment where screen size and resolution are less important (or loosely defined by the market for their software), mobile app developers have far less flexibility. How can you ensure that users will have a great experience with your app when it's impossible to optimize and may be running on widely varying hardware? It's the rare app developer that specifies system requirements other than OS version. The market and sales simply follow a different model that plays quite well on Apple's homogeneous hardware but just may need some tweaking in and Android world.
How can the app model adapt to a growing cadre of devices as Android as a platform explodes? There may have to be some customization of high-profit apps for the most popular devices or increased abilities to switch resolutions, frame rates, or other performance elements based on a user's device. There may also need to be further segmentation and fragmentation of the market, but this is hardly a desired outcome.
More likely, a model where system requirements become the norm again will be the best solution. While independent developers will still be able to crank out all the apps they can write, it would behoove device manufacturers, carriers, and those dealing in vertical markets (healthcare, for instance) to develop and optimize for whatever hardware platform they want to sell. I'd love to be able to search the Android Market for Droid-Incredible optimized apps and Motorola would sell even more Droids if they developed a killer app that ran only on the X with its speedy processor and giant screen.
Fragmentation isn't exactly fixing itself, but the consumer market is doing a decent job of solving the problem. The next big questions will be "What are the killer apps?", "Who will develop them?", and "Where will they run?"